Human monoclonal antibodies (mAb) are increasingly being used and developed for therapy and diagnosis in treating cancer, infection, and many inflammatory diseases. Despite recent advances in recombinant DNA technology, multiplexed screening tools, and fluid handling robots, it is a long and laborious process to identifying desired mAbs from human subjects. Herein, we propose a strategy for pairing cognate immunoglobulin heavy and light chain genes from the mixture of human B cells for unbiased, global mining of 'natural' antibody repertoires. Subsequently, human antibody libraries will be expressed in a system called, the 'yeast surface 2-hybrid', which we developed for the expression of a pair of proteins for interaction and discovery studies. Full-length antibodies expressed on the surface of yeast can be released by enzymatic proteolysis, which will facilitate downstream assays requiring soluble form of antibodies. Specific aims for the proposed study are 1) Devise a method for the pairing of cognate IgG heavy and light chain genes from a mixture of antibody producing cells; 2) Validate the platform for global, unbiased display and panning of natural human antibody repertoires. We will perform massively parallel sequencing to evaluate the diversity and any bias in our antibody library. Furthermore, we will examine the capacity of the proposed platform to isolate mAbs against selected common vaccine antigens (Pertussis, Hepatitis B virus).